> The same timing for frozenset((1,2,3)) (.29) is faster than the best
> timing for frozenset({1,2,3}), (.36),

I don't see the tuple form used anywhere in the code.
The timing is a bit quicker for the tuple form because the peephole optimizer constant folds the tuple (use dis to see this).


> so I will not change that 
> unless discussed and agreed on.

Maybe, I should just make the patch.  It's becoming harder to talk about than to just fix.